Foundations of Retrieval- Augmented Generation (RAG)

Training content

An accessible introduction to Large Language Models (LLMs) and Retrieval-Augmented Generation (RAG). Learn how modern RAG systems work and understand their core components, limitations and design choices.

This training is part of a three-day RAG course series. Day 1 focuses on conceptual foundations and can be attended independently, while Days 2 and 3 form a consecutive, hands-on technical module. Participants may register for Day 1 only, for the technical Days 2–3 bundle, or for the complete three-day program. Participants registered for the full 3-day program receive a €50 discount. To register for the complete program, register here and add-on days 2 and 3 in the checkout process.

This training will be conducted in English.

Program

Morning – LLM & RAG Fundamentals

• How LLMs work (tokens, prediction, hallucinations)
• Context window limitations and temperature
• Local vs API-based models (cost, privacy, performance)
• Fine-tuning vs RAG
• RAG principles and architectures

Afternoon – Core RAG Building Blocks

• Document processing (PDFs, text extraction)
• Chunking strategies and metadata
• Embeddings (local vs API)
• Vector storage and retrieval
• Dense vs sparse search (FAISS, BM25)
• Prompt structure and context management

Target audience

• IT services
• Public administrations
• Policy, innovation and digital teams
• Anyone working with or supervising AI solutions